lab10-roni-shen

Mobile Food Schedule Data

library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(stringr)
library(plotly)

## Loading required package: ggplot2

## 
## Attaching package: 'plotly'

## The following object is masked from 'package:ggplot2':
## 
##     last_plot

## The following object is masked from 'package:stats':
## 
##     filter

## The following object is masked from 'package:graphics':
## 
##     layout

# getting the data
github <- "https://raw.githubusercontent.com/ucb-stat133/stat133-fall-2018/master/"
datafile <- "data/mobile-food-sf.csv"
download.file(paste0(github, datafile), destfile = "mobile-food-sf.csv")

# importing data into R
dat <- read.csv('mobile-food-sf.csv', stringsAsFactors = FALSE)

Plots with “plotly”

# frequency of DayOfWeekStr in dat 
day_freqs <- table(dat$DayOfWeekStr)

barplot(day_freqs, border = NA, las = 3)

# plotting with plotly
plot_ly(x = names(day_freqs), 
        y = day_freqs,
        type = 'bar')

# day frequencies table
day_counts <- dat %>% 
  select(DayOfWeekStr) %>%
  group_by(DayOfWeekStr) %>%
  summarise(count = n()) %>%
  arrange(desc(count))

day_counts

## # A tibble: 7 x 2
##   DayOfWeekStr count
##   <chr>        <int>
## 1 Friday        1105
## 2 Wednesday     1095
## 3 Thursday      1090
## 4 Tuesday       1081
## 5 Monday        1080
## 6 Saturday       533
## 7 Sunday         263

# using poly_ly to plot a dataframe
plot_ly(day_counts, 
        x = ~DayOfWeekStr, 
        y = ~count,
        type = 'bar')

# reordering the bars by count
plot_ly(day_counts, 
        x = ~reorder(DayOfWeekStr, count), 
        y = ~count,
        type = 'bar')

Changing Times

# toy string
time1 <- '10AM'

# hour
str_sub(time1, start = 1, end = 2)

## [1] "10"

# period
str_sub(time1, start = 3, end = 4)

## [1] "AM"

times <- c('12PM', '10AM', '9AM', '8AM', '2PM')

# subset time
str_sub(times, start = 1, end = -3)

## [1] "12" "10" "9"  "8"  "2"

# subset period
str_sub(times, start = -2, end = nchar(times))

## [1] "PM" "AM" "AM" "AM" "PM"

# period
str_sub(times, start = -2)

## [1] "PM" "AM" "AM" "AM" "PM"

Function str_replace()

# replace AM/PM with ''
str_replace(times, pattern = 'AM|PM', replacement = '')

## [1] "12" "10" "9"  "8"  "2"

# Using times, create a numeric vector hours containing just the number time (i.e. hour)
hours <-  str_replace(times, pattern = 'AM|PM', replacement = '')

# Using times, create a character vector periods containing the period, e.g. AM or PM
periods <- str_sub(times, start = -2)

# Use plot_ly() to make a barchart of the counts for AM and PM values.
periods_count = table(periods)
plot_ly(x = names(periods_count),
        y = periods_count,
        type = 'bar')

# Write R code to create a vector start24 that contains the hour in 24hr scale.

#' @title convert to 24
#' @description coverts a character vector of 12hr cycle times to numeric vector of 24 hr cycle times
#' @param x a character vector of 12hr cycle times
#' @return a numeric vector of 24 hr cycle times
convertto24 <- function (x) {
  y <- rep("", length(x))
  x_hours <- str_sub(x, start = 1, end = -3)
  x_periods <- str_sub(x, start = -2)
  for (i in 1: length(x)) {
    if (x_periods[i] == 'PM') {
      y[i] = strtoi(x_hours[i]) + 12
    } else {
      y[i] = strtoi(x_hours[i])
    }
  }
  return(as.numeric(y))
}

start24 = convertto24(times)

# Add two columns start and end to the data frame dat, containing the starting and ending hour respectively (columns must be "numeric").
dat$start <- convertto24(dat$starttime)
dat$end <- convertto24(dat$endtime)

# With the starting and ending hours, calculate the duration, and add one more column duration to the data frame dat
dat$duration <- dat$end - dat$start 

Latitude and Longitude Coordinates

# location with longitude and latitude
loc1 <- "(37.7651967350509,-122.416451692902)"

# "remove" opening parenthesis 
str_replace(loc1, pattern = '\\(', replacement = '')

## [1] "37.7651967350509,-122.416451692902)"

# "remove" closing parenthesis
str_replace(loc1, pattern = '\\)', replacement = '')

## [1] "(37.7651967350509,-122.416451692902"

# removing both parenthesis: str_replace only replaces one
str_replace(loc1, pattern = '\\(|\\)', replacement = '')

## [1] "37.7651967350509,-122.416451692902)"

# use str_replace_all to replace all 
str_replace_all(loc1, pattern = '\\(|\\)', replacement = '')

## [1] "37.7651967350509,-122.416451692902"

lat_lon <- str_replace_all(loc1, pattern = '\\(|\\)', replacement = '')

# getting ride of commas
str_replace(lat_lon, pattern = ',', replacement = '')

## [1] "37.7651967350509-122.416451692902"

# string split in stringr
str_split(lat_lon, pattern = ',') # note: str_split returns a list 

## [[1]]
## [1] "37.7651967350509"  "-122.416451692902"

# vector with location values
locs <- c(
  "(37.7651967350509,-122.416451692902)",
  "(37.7907890558203,-122.402273431333)",
  "(37.7111991003088,-122.394693339395)",
  "(37.7773000262759,-122.394812784799)",
  NA
)

# list containing longitude and latitude
lat_lon2 <- str_replace_all(locs, pattern = '\\(|\\)', replacement = '') %>% str_split(pattern = ',')

# getting just the lat
lat <- lapply(lat_lon2, function(x) x[1])

# getting just the lon
lon <- lapply(lat_lon2, function(x) x[2])

# convert from list to vector
lat <- as.numeric(unlist(lat))
lon <- as.numeric(unlist(lon))

# adding lat and lon to dat
dat$lat <- as.numeric(unlist(dat$Location %>% str_replace_all(pattern = '\\(|\\)', replacement = '') %>% str_split(pattern = ',') %>% lapply(function(x) x[1])))
dat$lon <- as.numeric(unlist(dat$Location %>% str_replace_all(pattern = '\\(|\\)', replacement = '') %>% str_split(pattern = ',') %>% lapply(function(x) x[2])))

Plotting locations on a map

# simple option
plot(dat$lon, dat$lat, pch = 19, col = "#77777744")

# Scatterplots with "plotly"
plot_ly(x = dat$lon, y = dat$lat)

## No trace type specified:
##   Based on info supplied, a 'scatter' trace seems appropriate.
##   Read more about this trace type -> https://plot.ly/r/reference/#scatter

## No scatter mode specifed:
##   Setting the mode to markers
##   Read more about this attribute -> https://plot.ly/r/reference/#scatter-mode

## Warning: Ignoring 40 observations

# default scatterplot
plot_ly(x = dat$lon, y = dat$lat, type = 'scatter', mode = 'markers')

## Warning: Ignoring 40 observations

# similar to ggplot
plot_ly(data = dat, x = ~lon, y = ~lat, type = 'scatter', mode = 'markers')

## Warning: Ignoring 40 observations

Maps with “RgoogleMaps”

# install.packages("RgoogleMaps")
library(RgoogleMaps)

# coordinates for center of the map
center <- c(mean(dat$lat, na.rm = TRUE), mean(dat$lon, na.rm = TRUE))

# zoom value
zoom <- min(MaxZoom(range(dat$lat, na.rm = TRUE), 
                    range(dat$lon, na.rm = TRUE)))

# san francisco map
map1 <- GetMap(center=center, zoom=zoom, destfile = "san-francisco.png")

PlotOnStaticMap(map1, dat$lat, dat$lon, col = "#ed4964", pch=20)

Maps with “ggmap”

# install.packages("ggmap")
library(ggmap)

## 
## Attaching package: 'ggmap'

## The following object is masked from 'package:plotly':
## 
##     wind

# let's get rid of rows with missing values
dat <- na.omit(dat)

# ggmap typically asks you for a zoom level, but we can try using ggmap's make_bbox function:
sbbox <- make_bbox(lon = dat$lon, lat = dat$lat, f = .1)
sbbox

##       left     bottom      right        top 
## -122.48867   37.69985 -122.36281   37.81595

# get a 'terrain' map
sf_map <- get_map(location = sbbox, maptype = "terrain", source = "google")

## Warning: bounding box given to google - spatial extent only approximate.

## converting bounding box to center/zoom specification. (experimental)

## Map from URL : http://maps.googleapis.com/maps/api/staticmap?center=37.757897,-122.425744&zoom=13&size=640x640&scale=2&maptype=terrain&language=en-EN&sensor=false

ggmap(sf_map) + 
  geom_point(data = dat, 
             mapping = aes(x = lon, y = lat), 
             color = "red", alpha = 0.2, size = 1)

## Warning: Removed 98 rows containing missing values (geom_point).

Let’s look for specific types of food

# dat contains a column optionaltext describing the types of food and meals served by the food trucks
dat$optionaltext[1:3]

## [1] "Tacos, Burritos, Tortas, Quesadillas, Mexican Drinks, Aguas Frescas"   
## [2] "Cold Truck: sandwiches, drinks, snacks, candy, hot coffee"             
## [3] "Cold Truck: Pre-packaged Sandwiches, Various Beverages, Salads, Snacks"

# first 10 elements
foods <- dat$optionaltext[1:10]

# Use str_detect() (or equivalently grep()) to match "Burritos" and "burritos".
str_detect(tolower(foods), "burritos")

##  [1]  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE

# If you use grepl(), you can use ignore.case = TRUE to match for both.
grepl("burritos", foods, ignore.case = TRUE)

##  [1]  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE

# Try another pattern: e.g. "tacos", or "quesadillas"
grepl("taco", foods, ignore.case = TRUE)

##  [1]  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE

# Now create a data frame burritos by subsetting (i.e. filtering) the data frame to get only those rows that match "burritos"
burritos <- dat[grepl("burritos", dat$optionaltext, ignore.case = TRUE),]

# Use the lat and lon corrdinates in burritos to display a map of locations with burritos (see map below).
plot_ly(data = burritos, x = ~lon, y = ~lat, type = 'scatter', mode = 'markers')

burritos = na.omit(burritos)
ggmap(sf_map) + 
  geom_point(data = burritos, 
             mapping = aes(x = lon, y = lat), 
             color = "red", alpha = 0.2, size = 1)

## Warning: Removed 18 rows containing missing values (geom_point).

# Experiment with other types of foods
ice_cream <- dat[grepl("ice cream", dat$optionaltext, ignore.case = TRUE),]
ice_cream = na.omit(ice_cream)
ggmap(sf_map) + 
  geom_point(data = ice_cream, 
             mapping = aes(x = lon, y = lat), 
             color = "red", alpha = 0.2, size = 1)

## Warning: Removed 17 rows containing missing values (geom_point).

salad <- dat[grepl("salad", dat$optionaltext, ignore.case = TRUE),]
salad = na.omit(salad)
ggmap(sf_map) + 
  geom_point(data = salad, 
             mapping = aes(x = lon, y = lat), 
             color = "red", alpha = 0.2, size = 1)

## Warning: Removed 24 rows containing missing values (geom_point).

noodle <- dat[grepl("noodle", dat$optionaltext, ignore.case = TRUE),]
noodle = na.omit(noodle)
ggmap(sf_map) + 
  geom_point(data = noodle, 
             mapping = aes(x = lon, y = lat), 
             color = "red", alpha = 0.2, size = 1)

## Warning: Removed 25 rows containing missing values (geom_point).

# Challenge: try use facetting to show a type of food per facet (e.g. one facet for burritos, another for quesadillas, another one for tacos, etc)
ggmap(sf_map) + 
  geom_point(data = dat, 
             mapping = aes(x = lon, y = lat), 
             color = "red", alpha = 0.2, size = 1) +
  facet_grid(~optionaltext)

## Warning: Removed 98 rows containing missing values (geom_point).